Just spent days tracking problems down only to find it was a bug in the
load tester we use at work -- it was including an extra CRLF after a
POST request.  Apparently there are quite a few buggy HTTP/1.0 clients
out there in this respect.

from preamble:

"Change Set:		Comanche-leading-crlf-robustness-jf
Date:			6 March 2004
Author:			Julian Fitzell

RFC 2068 says in section 4.1 (Message Types) that 'In the interest of
robustness, servers SHOULD ignore any empty line(s) received where a
Request-Line is expected'.

This changeset adds a method to HttpRequest that pulls out the first
*non-empty* line for use as the Status/Request line."!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Comanche-leading-crlf-robustness-jf.cs.gz
Type: application/octet-stream
Size: 662 bytes
Desc: not available
Url : http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20040306/26f582c4/Comanche-leading-crlf-robustness-jf.cs.obj